Daily Digest: OpenAI safety (and drama...again)

Want to get in front of 100k AI enthusiasts? Work with us here

Hello folks, here’s what we have today;

PICKS

1. New tutorial: Create a sales prospecting AI copilot (without code)

   • ICYMI: Ben’s Bites Pro will go from $150 to $250 (no subscription) in June (next month), so if you’ve been thinking about it and want to lock in the lower price, now’s the time to do it! Sign up here.

2. OpenAI makes a new safety committee and hints at new models. Sam Altman is one of the leads for this new safety and security committee. According to the announcement, OpenAI has recently begun training its next frontier model and expects it to unlock new capabilities.🍿Our Summary (also below)

3. Helen Toner, a former OpenAI board member, is speaking out. She claims the board wasn't informed about ChatGPT by Sam Altman but instead found out through Twitter. She even accuses Altman of "outright lying" at times, keeping the board in the dark about OpenAI's projects and safety efforts.

4. Gemini 1.5 Pro ranks #2 on the LMSYS leaderboard with 1268 Elo. Gemini Advanced (Elo: 1267) is just behind it at #3. Google is quickly closing the gap between its models and OpenAI’s GPT-4o (Elo: 1287).

from our sponsor

Aethir: On-demand H100s for $1.99/h - Stop waiting, start creating!

Zuck stealing all your GPUs? Rent on-demand HGX H100 GPUs for just $1.99/hour. No contracts, no pre-payments, be up and running in 15 min. Short-term & no-term options + Large local clusters. We make GPUs more accessible, so you can build faster!

Include "AethirBen" in your Message for access to this exclusive price! Access Here

TOP TOOLS

• Roadway - The first growth marketing analytics and automation platform.

• Frontly - Build AI-powered SaaS apps and internal tools with no code.

• Atlabs - Your AI social video co-pilot.

• Lawformer AI - Break complex contracts into smaller elements.

• Reforge Extension - An expert by your side in the places you work.

• Skej - Your new AI scheduling assistant.

• Tonic Textual - Get your unstructured data AI-ready in minutes.

• HyperAide - A simple, AI first task manager.

• FlowTunes - Endless AI-generated music for focus.

• Autospec - Open-source agent that generates E2E tests for your web app.

• HuggingFace Chat gets access to tools starting with Cohere’s Command R model.

View more →

NEWS

• Jan Leike joins Anthropic as the new safety head weeks after resigning as OpenAI’s Superaliment co-lead. 👀 possible org changes at Anthropic.

• Skej’s AI meeting scheduling assistant works like adding an EA to your email

• SWE-agent - Agent-computer interfaces enable automated software engineering.

• Innovation Laboratory - A 4-week guided summer course on how to train your own AI model by HUG & Stability AI.

• Nvidia’s cloud spending soars as it looks beyond chips.

• Google adds AI-powered features to Chromebook and offers a free year of AI premium plan.

• China’s $47B semiconductor fund puts chip sovereignty front and centre.

• What does the public in six countries think of generative AI?

• What we learned from a year of building with LLMs - Part 1.

View more →

QUICK BITES

OpenAI just announced a new Safety and Security Committee led by Sam Altman, Bret Taylor, and other board members. This move comes hot on the heels of some top scientists leaving, including those leading the "superalignment" team focused on tackling long-term AI risks.

What is going on here?

OpenAI makes a new safety committee and hints at new models.

What does this mean?

The committee, a group of experts including the CEO and other board members, will take 90 days to look closely at and strengthen the safety measures they already have in place. This is all happening as OpenAI starts training its next big AI model, which they think will be a major step up in what AI can do.

Besides the board members, the committee also includes some of OpenAI's top experts in tech and policy, like Aleksander Madry, Lilian Weng, John Schulman, Matt Knight, and Jakub Pachocki. They'll also be getting advice from former government officials like Rob Joyce and John Carlin, who know a thing or two about cybersecurity and national security.

Why should I care?

OpenAI is a leader in AI, so their choices about safety could influence other companies in the field. However, some former employees have criticized OpenAI, and there are worries about how committed they are to safety after losing some key members of their safety team.

I want more than “iterative deployment is the solution” from the committee's findings. Let’s see how OpenAI plans to address these concerns and prioritize safety as its tech advances.

Share this story

Ben’s Bites Insights

We have 2 databases that are updated daily which you can access by sharing Ben’s Bites using the link below;

• All 10k+ links we’ve covered, easily filterable (1 referral)

• 6k+ AI company funding rounds from Jan 2022, including investors, amounts, stage etc (3 referrals)